A Three-Color Coupled Level-Set Algorithm for Simultaneous Multiple Cell Segmentation and Tracking

摘要: High content computational analysis of time-lapse microscopic cell images requires accurate and efficient segmentation tracking. In this work, we introduce “3LS”, an algorithm using only three level sets to segment track arbitrary number cells in images. The positions are determined the first frame by extracting concave points fitting ellipses after initial segmentation. We construct a graph representing background with vertices their adjacency relationships edges. Each vertex is assigned color tag applying coloring algorithm. way, boundary each can be embedded one set functions. “3LS” implemented existing coupled active contour framework (nLS) [1] handle overlapped during However, improve nLS new volume conservation constraint (VCC) prevent shrinkage or expansion on whole boundaries produce more tracking touching cells. When tested four different image sequences, 3LS outperforms original other relevant state-of-the-art counterparts both however notable reduction time.